[Ifeffit] Vanadium-Chromium Prussian Blue Analog XANES fitting

[Adora Graham]

Matt,

"It can be challenging to extrapolate from XANES with different ligands,
and Prussian Blue with linear CN ligands might act quite differently from O
ligands.   Like, I wouldn't necessarily suspect the VS04 was a particularly
good reference for you, though I could be wrong."

I think this is likely the case as well, seeing as I am getting
unreasonable results from the oxide standard extrapolation. Is there any
other way I could use what I have to extrapolate the oxidation state from
my data?


I talked to the person in charge of the XRF and other equipment in our
nanofab facility and had been advised that the XRF that we have available
will not be accurate enough for determining V to Cr ratios and that we
would need a Wavelength Dispersive XRF or equivalent to get quantitative
results. We have tried ICP-MS, but encountered issues with the facility
that ran our samples. It might be worth trying again to see if we can find
a different facility to run these measurements for us, as that should give
reasonable values for elemental composition and we can then infer the
oxidation state from a charge balance perspective. We had tried XPS, but
have some concerns as to whether or not the sample composition will  be the
same at the surface compared to the bulk.

The samples were prepared by grinding the compounds with boron nitride
using a mortar and pestle and sealing them within Kapton tape (using a note
card as the window). With this setup, I don't think it will be easy, if at
all possible to deduce the thickness. Since the edges are so close, the
intention was to run both edges together on the same scan, but we were
advised to run them separately by the technician at Argon. As far as I know
(I did not run the scans personally) the samples were not moved or changed
between the scans, so at least we should be looking at the same part of the
sample for both energy edges.

Thanks,
Adora

On Thu, Jul 30, 2015 at 8:51 PM, Matt Newville <newville at cars.uchicago.edu>
wrote:

> Adora,
>
> On Thu, Jul 30, 2015 at 6:38 PM, Adora Graham <abaldwin2242 at gmail.com>
> wrote:
>
>> Hello,
>>
>> First of all, I am still very new to XAS analysis, so please excuse any
>> misunderstandings I may have.
>>
>> I have a series of compounds nominally consisting of KxVx[Cr(CN)6]
>> Prussian blue analogs (where x is ~1) with Cr(III) and presumably V(II)
>> present in the system. From other evidence, the V(II) may be oxidizing to
>> V(III). I have XAS data collected at the V and Cr K-edges from APS and have
>> been struggling through trying to analyze my data using Athena.
>>
>> I am trying to get two things out of this data:
>>
>> First, I would like to get the oxidation state of the Vanadium sites. If
>> they are not V(II), to get the ratio of V(II) to V(III) present, if
>> possible. I have been trying to follow a similar analysis to Wong (Phys.
>> Rev. B, 1984, 30, 10, 5596-5610). However, I seem to be running into the
>> issue of finding standards with a similar ligand set to my sample. I have
>> XAS data for the following standards: VO, V2O3, V2O4, V2O5, V(acac)3,
>> VO(acac), VOSO4, and VSO4. I start with VSO4 as my initial vanadium source,
>> so that could be useful for a comparison. Our initial idea was to use the
>> vanadium oxides to give a standard curve to then extrapolate the oxidation
>> state of my compound, but using this gives me an oxidation state of ~4.2
>> which seems unreasonable for my system. I am kind of lost otherwise as to
>> what else I could do with this.
>>
>>
> From a quick look at the V spectrum, it looks much more reduced than 4+ to
> me -- more reduced than 3+, even.   At least when compared to the metal
> oxide series.  How were you doing the comparison?
>
> It can be challenging to extrapolate from XANES with different ligands,
> and Prussian Blue with linear CN ligands might act quite differently from O
> ligands.   Like, I wouldn't necessarily suspect the VS04 was a particularly
> good reference for you, though I could be wrong.
>
>
> Second, I would like to get the ratio of V to Cr. I was told I could
>> compare the un-normalized K-edge heights of the vanadium to that of the
>> chromium to get this ratio. This doesn't seem right, as I would think that
>> there should be some relative intensity factor involved. Also, this gives
>> me a fairly unreasonable value of 4:1, since XPS measurements give me
>> values of around 1:1.
>>
>> Can you point me in the right direction for these issues?
>>
>
> Generally XRF or mass spec (if surface sensitive is OK XPS)  is preferred
> for elemental concentrations over XAFS edge jumps. To use XAFS edge steps,
> you'd have to know the thickness of the sample to high precision, and ought
> to correct for everything else in the beam.   Transferring that between
> different edges would be another complication, though V to Cr is pretty
> close.   Were the edge jump data on different edges of the sample piece of
> sample?
>
>
> --Matt
>
>
>
> _______________________________________________
> Ifeffit mailing list
> Ifeffit at millenia.cars.aps.anl.gov
> http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://millenia.cars.aps.anl.gov/pipermail/ifeffit/attachments/20150730/3b1493b0/attachment.html>